Applying of Forecast-Revise EKF Algorithm in Autonomous Navigation System

Gong-yuan ZHANG,Yong-mei CHENG,Feng YANG,Quan PAN,Cong GU
DOI: https://doi.org/10.3873/j.issn.1000-1328.2009.06.029
2009-01-01
Abstract:In order to improve the nonlinear estimation accuracy in autonomous navigation of spacecraft, via analyzing the drawbacks of linearization in EKF(Extended Kalman Filter), the influence of the difference expansion point in linearization of non-linear function and the difference valued point in Jacobin matrix to the accuracy of" linearization was studied based on the theory of mathematics analyzing. Then the linearization expansion way of generic EKF was improved, and a new algorithm called Forecast-Revise EKF (FR-EKF) based on mean value theorem and the smoothing estimation was proposed. Simulation result shows that the new algorithm is prior to EKF regard to accuracy.
What problem does this paper attempt to address?